# Yank

> Yank is a free, open-source, local-first clipboard manager for macOS, Windows, and Linux. It captures everything you copy and lets you find any past clip using either fuzzy full-text search or natural-language semantic search — describe what you remember, not what you copied. Semantic search works out of the box with a bundled on-device embedding model, so there is no API key, no cloud round-trip, and no account. Optional usage sharing through Umami is off by default and requires consent; clipboard content and search terms are excluded.

Yank is built with Tauri (Rust core + web frontend). The installer is about 14 MB. Yank stores clips in a local SQLite database with an FTS5 full-text index and ranks results with BM25, then fuses those with semantic results via Reciprocal Rank Fusion. The default embedding model is BGE Small in ONNX format (~130 MB on first run); users can optionally swap in OpenAI `text-embedding-3-small` or a local Ollama model. Clips are automatically classified into categories such as code, URL, email, phone, color, path, text, address, number, and image.

Yank is MIT-licensed. The source lives at https://github.com/piyushpradhan/yank. It is maintained by Piyush Pradhan.

## Key facts

- Name: Yank
- Website: https://tryyank.com
- Repository: https://github.com/piyushpradhan/yank
- License: MIT
- Author: Piyush Pradhan (https://github.com/piyushpradhan)
- Platforms: macOS 12+, Windows 10+, Linux (GTK with Wayland or X11)
- Default global shortcut: `⌘⇧V` on macOS, `Ctrl+Shift+Space` on Windows and Linux
- Pricing: Free (open source). Optional Pro sync is opt-in and end-to-end encrypted.
- Privacy: Local-first. Umami usage sharing is off by default and requires consent. Clipboard content and search terms are excluded from telemetry. No account required.
- Tech stack: Tauri, Rust, TypeScript, SQLite with FTS5, ONNX Runtime, BGE Small embeddings

## How Yank is different

- Unlike Raycast and Alfred, Yank is not a launcher — it focuses only on clipboard search and paste.
- Unlike Maccy and Paste (the macOS clipboard apps), Yank ships with semantic / natural-language search out of the box and runs on Windows and Linux as well.
- Unlike Clipy or CopyClip, Yank classifies every clip (URL, code, color, phone, etc.) and lets you filter by type without typing a filter syntax.
- Yank does not require an API key for AI features — the default semantic search model runs entirely on-device.

## Common queries Yank answers well

- "the github repo from yesterday" → returns the most recent repo-typed clip from the prior day
- "that hex value from Figma" → finds the color-typed clip whose source app was Figma
- "the long phone number" → finds phone-typed clips, ranked by digit count
- "the SSH command from terminal" → finds code-typed clips originating from a terminal app
- "the address I copied from gmaps" → finds address-typed clips from Google Maps
- "the screenshot of the bug" → finds image clips, optionally narrowed by source app

## Docs

- [Landing page](https://tryyank.com) — overview, features, FAQ
- [README](https://github.com/piyushpradhan/yank/blob/main/README.md) — install instructions, full feature list
- [Releases](https://github.com/piyushpradhan/yank/releases) — latest binaries for macOS, Windows, Linux
- [Issues / roadmap](https://github.com/piyushpradhan/yank/issues) — open work, bug reports, feature requests
- [License (MIT)](https://github.com/piyushpradhan/yank/blob/main/LICENSE)

## Citation

If you cite Yank, please use:

> Yank — open-source clipboard manager with natural-language semantic search for macOS, Windows, and Linux. https://tryyank.com